2
$\begingroup$

Let $T_n$ and $T$ be linear maps from Banach space $X$ to a Banach space $Y$.

Suppose $T_n$ satisfies $T_nx \to Tx$ (convergence in the $Y$ norm) for all $x \in X$.

Let $x_n \rightharpoonup x$ in $X$ (weak convergence). Under what conditions on $T_n$ and $T$ does this imply that $$T_nx_n \rightharpoonup Tx?$$

Let us assume all maps are bounded if necessary.

What if instead of $x_n \rightharpoonup x$ we hypothesised $x_n \to x$?

$\endgroup$

migrated from mathoverflow.net Apr 21 '15 at 19:50

This question came from our site for professional mathematicians.

  • 1
    $\begingroup$ If the maps aren't continuous (bounded), you can't expect (weak) convergence. If we assume boundedness and strong convergence $x_n \to x$, the Banach-Steinhaus theorem implies the result. (Edit: added requirement for strong convergence.) $\endgroup$ – Daniel Fischer Apr 21 '15 at 22:08
  • $\begingroup$ @DanielFischer: How do you get this from Banach-Steinhaus? I'm not seeing it. $\endgroup$ – Nate Eldredge Apr 22 '15 at 0:47
  • $\begingroup$ @NateEldredge Good thing you don't see it, because it doesn't hold if we have only weak convergence $x_n \rightharpoonup x$. (Counterexample: $X = \ell^p(\mathbb{N}),\; 1 < p < \infty$ with $x_n = e_n$ and $T_n(y) = y_n$.) I can't reconstruct what mistake made me think otherwise yesterday, unfortunately. Thanks for the heads-up. $\endgroup$ – Daniel Fischer Apr 22 '15 at 8:37
  • $\begingroup$ @DanielFischer Can you give a hint why it implies the result? $\endgroup$ – C_Al Apr 23 '15 at 14:59
  • 1
    $\begingroup$ Yes, Banach-Steinhaus gives uniform boundedness of the $T_n$, which is equicontinuity, and that gives norm convergence of $T_n x_n$ to $Tx$ if we assume $x_n \to x$. If we assume only $x_n \rightharpoonup x$, we would need norm convergence $T_n \to T$ to be able to deduce $T_n x_n \rightharpoonup Tx$. $\endgroup$ – Daniel Fischer Apr 23 '15 at 15:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy